lable at ScienceDirect

Journal of Power Sources 365 (2017) 92e97

Contents lists avai

Journal of Power Sources

journal homepage: www.elsevier .com/locate/ jpowsour

Short communication

Novel structure design of composite proton exchange membraneswith continuous and through-membrane proton-conducting channels

Hang Wang a, b, Chenxiao Tang b, Xupin Zhuang a, b, *, Bowen Cheng a, **, Wei Wang a,Weimin Kang a, b, Hongjun Li a

a State Key Laboratory of Separation Membranes and Membrane Processes, Tianjin Polytechnic University, Tianjin, 300387, PR Chinab College of Textile, Tianjin Polytechnic University, Tianjin, 300387, PR China

h i g h l i g h t s

* Corresponding author. State Key Laboratory ofMembrane Processes, Tianjin Polytechnic University,** Corresponding author.

E-mail addresses: zhxupin@139.com (X. Zhu(B. Cheng).

http://dx.doi.org/10.1016/j.jpowsour.2017.08.0720378-7753/© 2017 Elsevier B.V. All rights reserved.

g r a p h i c a l a b s t r a c t

� Novel PEMs with through-membraneproton-conducting channels areproposed.

� The structured PEMs were preparedvia solution blowing and hot-pressedtreatment.

� The effects of nanofiber distributionon the properties of PEMs wereinvestigated.

pm.co

m.cn

a r t i c l e i n f o

Article history:Received 25 February 2017Received in revised form1 August 2017Accepted 18 August 2017

Keywords:Through-membraneProton exchange membraneHot-pressSolution blowingPhenolphthalein sulfonated poly(ethersulfone)Poly(vinylidene fluoride)

www.sa b s t r a c t

The primary goal of this study is to develop a high-performanced proton exchange membrane with thecharacteristics of through-membrane and continuous solution blown nanofibers as proton-conductingchannels. The curled sulfonated phenolphthalein poly (ether sulfone) and poly (vinylidene fluoride)nanofibers were separately fabricated through the solution blowing process which is a new nanofiberfabricating method with high productivity, then they were fabricated into a sandwich-structured mat.Then this sandwich-structured mat was hot-pressed to form the designed structure using differentmelting temperatures of the two polymers by melting and making poly (vinylidene fluoride) flow intothe phenolphthalein poly (ether sulfone) nanofiber mat. The characteristics of the composite membrane,such as morphology and performance of the membrane, were investigated. The characterization resultsproved the successful preparation of the membrane structure. Performance results showed that thenovel structured membrane with through-membrane nanofibers significantly improved water swellingand methanol permeability, though its conductivity is lower than that of Nafion, the cell performanceshowed comparable results. Therefore, the novel structure design can be considered as a promisingmethod for preparing of proton exchange membranes.

© 2017 Elsevier B.V. All rights reserved.

Separation Membranes andTianjin, 300387, PR China.

ang), bowen15@tjpu.edu.cn

1. Introduction

At present, energy and environmental problems are the focusesof global concern. Direct methanol fuel cells (DMFCs) are consid-ered to be a promising technology, because of their clean and

m

nFig. 1. Schematic representations of the structure of the designed membranes.

H. Wang et al. / Journal of Power Sources 365 (2017) 92e97 93

www.sp

efficient power sources [1]. As the core component of DMFCs,proton exchange membranes (PEMs) require superior proton con-ductivity, lowmethanol permeability and chemical stability [2]. Forcurrent DMFCs, PEMs comprising perfluorosulfonic acid polymers,such as Nafion, are the material of choice; however, per-fluorosulfonic acid polymers show limited device lifetime causedby high methanol permeability and high swelling. Therefore, re-searchers have become motivated to overcome these issues and toimprove PEM performances.

Extensive efforts were made to make some progress in someaspects. These efforts include (i) optimizing polymer structure[3e6], (ii) developing non-fluorinated ion conducting membranesto replace Nafion [2], and (iii) preparing blend membranes withvarious materials [7,8]. In recent years, designing stable PEMschemically and mechanically while maintaining high ionic con-ductivity has been a requirement [3]. The development of nanofiber(NF) fabricating technology has resulted in the increase in thenumber of studies that focus on impregnating NFs into the polymermatrix to obtain high-performance PEMs; in this approach, NFs,which play the role of long channels for proton conduction, showbetter advantages than particle and rod-like additives because oftheir large lengthediameter ratio [9,10]. It has been proven thatprotons could transport efficiently through hydrophilic ionic clus-ters along NFs through hopping, vehicular and surface transport innanofiber composite membranes (NCMs) [11,12]. NFs decrease themethanol permeability by blocking methanol transmissionpathway in the composite membranes as a methanol barrier layer[9]. The NFs in NCMs can form a three-dimensional networkstructure that has a good dimensional stability, and the swelling ofmembranes can be suppressed [13]. Many researchers [14,15]introduced different kinds of NFs into a matrix to fabricate mem-branes with nanofibrous skeleton, which showed propertyenhancement for composite membranes. Therefore, NFs play amajor and positive role in PEMs [16e18].

According to the review of the related literature, the majority ofthese works were conducted by applying electrospun NFs to rein-force PEMs with impregnation [19,20]. Electrospinning is the mainprocess used to prepare fibrous mat continuously with a highsurface area and remarkable porosity [15,21]. However, the naturalcharacteristics of the electrospinning process make NFs combinetightly to form a dense NFs network, which result in poor pro-cessability when NFs are used to reinforce PEMs through theimpregnation process. The obtained composite membranes easilysuffer from internal defects with some voids within compositestructures [22]. Another serious structural defect is the unavoidablepure polymer layers at the surfaces of composite membranes[23,24]. This phenomenon indicates that the surface layers do notcontain any NF, and these layers act as a “barrier layer” in protontransportation from one side to another side of PEMs. Thus, thedistribution of NFs along the direction of membrane thickness mustbe improved.

In this work, we proposed a novel structure, which is shown inFig. 1. We designed the NFs to build the transmembrane protonconduction path by distributing through PEMs; thus, protons couldbe transferred effectively and directly from one side to another side.Solution blowing process (SBP), which is a novel NF fabricatingmethod, could be applied to achieve the proposed structure. In thisprocess, polymer solutions are blown to ultrathin solution streamsby a high-speed gas flow and solidify to NFs after solvents evapo-ration [25e29]. Unlike electrospun NFs, solution-blown NFs arecommonly curled in three dimensions; this feature results in aloose construction [27,28]. Otherwise, a large number of crimpedfibers are closely entangled to form a stable and entangled struc-ture. Two different polymer NF mats were fabricated separatelythrough SBP. Then, these polymer NF mats were constructed as a

.com

.csandwiched structure. When an obvious difference existed in themelting temperatures of the two polymers and when the mat washot-pressed over a low melting temperature, this kind of bicom-ponent NF mat could be converted to a dense membrane with curlyNFs distributed, as designed in Fig. 1. In this study, hot-presstreatment was used to melt polymer with a low melting temper-ature and to make this polymer flow into the void spaces betweenNFs with a high melting temperature. In brief, a novel PEM wasdesigned by solution blowing of the two polymers with distinctmelting temperature into the bicomponent NFmat, which was hot-pressed into a dense membrane with curly transmembrane NFs asproton conduction path.

With respect to polymer choice, sulfonated phenolphthaleinpoly (ether sulfone) (SPES-C) was selected as the high-meltingpoint polymer. In recent years, sulfonated aromatic hydrocarbonpolymers have been the focus of the development of novel polymerelectrolyte membranes, such as sulfonated poly (ether ether ke-tone) (SPEEK) [30], sulfonated poly (ether sulfone) (SPES) [31], andsulfonated polyphenylsulfone (SPPSU) [32]. Phenolphthalein poly(ether sulfone) (PES-C) is a novel thermoplastic engineering ma-terial with phenolphthalein side chains. Thus far, PES-C has beenthe least reported among the various aromatic hydrocarbon poly-mers. SPES-C membranes have lower methanol permeability andhigher resistance to swelling than conventional SPES membranesbecause of the rigidity of the main chains and the steric hindranceeffect of phenolphthalein side chains [33]. Poly (vinylidene fluo-ride) (PVDF), which has a melting temperature of 172 �C, is a kind oflow-melting point polymer that meets our structural design re-quirements. PVDF can be also used as melting components.Moreover, PVDF, which is a kind of hydrophobic material that canreduce methanol transport rate in PEMs [34], is commonly appliedfor fuel cells [35]. The PVDF with melting temperature lower thanthat of SPES-C, which is selected as low-melting point polymer andthe methanol-resistant layer in this work. Furthermore, it can beused to foil the effects of through-membrane proton channelsowing to the lack of ionic exchange capacity, and verify the suc-cessful construction of the structures of through-membrane pro-ton-conducting channels. PVDF and SPES-C NF mats were firstfabricated using the SBP separately. Then PVDF and SPES-C matswere made to a sandwiched structure, then compacted with

H. Wang et al. / Journal of Power Sources 365 (2017) 92e9794

heating (180 �C, up to melting temperature of PVDF) to obtainmembranes by allowing molten PVDF to flow into the void space.The rationale for this study was twofold: (i) We sought to designdense NCMs with through-membrane NFs, proton-conductingchannels, and methanol-blocking layer. (ii) We sought to under-stand the effects of through-membrane NFs on the final membraneproperties. The proton conductivity, swelling ratio, permeability ofmethanol, and mechanical property of PEMs were investigated.

We believe that hot-press NFs mat treatment was developed inthis study as a route to prepare PEMs for the first time. This treat-ment simplifies the membrane fabrication procedure.

2. Experimental sections

2.1. Materials

PVDF was purchased from Solvay Co. PES was purchased fromChangchun Institute of Applied Chemistry. N,N-dimethylforma-mide was procured from Tianjin Kermel Chemical Reagent Co. Ltd.All chemicals were of analytical grade.

m

www.sp2.2. Preparation of composite membrane

SBP is a new NF fabricating method with high productivity, andthe experimental setup used to prepare nonwoven NFs was shownin our previous work [25]. The detailed synthetic method of SPES-Cis shown in the supporting information (SI). Two polymer spinningsolutions were prepared usingN,N-dimethylformamide as solvents.PVDF and SPES-C NF mats were first fabricated through SBP sepa-rately. PVDF and SPES-C mats were constructed as a sandwichedstructure composed of outer NF layers of SPES and a central basePVDF polymer layer. Then, these mats were preheated in a hotpresser (CB-950Z5 hot-presser, China) at 180 �C without pressure.After 5 min, the multilayer mats were compacted by heating at180 �C and a pressure of 10MPa for 10min. During hot pressing, themolten PVDF flowed into the void spaces between SPES-C NFs tocreate a fully dense membrane structure. Finally, the membranewas annealed at 180 �C. The weight-based compositions of PEMs(SPES-C/PVDF) were 80/20, 70/30, and 60/40 and were labeled asSPES-C/PVDF-20, SPES-C/PVDF-30, and SPES-C/PVDF-40, respec-tively. In addition, the membrane with pure PVDF on the surfaceand SPES-C NFs in the membrane was prepared for a comparison ofconductivity and was labeled as SPES-C/PVDF-F. The thicknesses ofall the composite membranes are approximately 150 mm.

2.3. Structural and morphological characterization of thenanofibers and composite membranes

The morphologies of the NFs and membranes were character-ized using a scanning electron microscope (SEM; Hitachi S-4800),and the atomic force microscope (AFM) images were recorded intapping mode (CSPM5500).

X-ray diffraction (XRD) measurements were performed via XRDspectroscopy (D8 Discover with GADDS).

2.4. Tensile strength

The tensile strength of dried membranes (dimensions:1 cm � 4 cm) was measured using an Instron universal testingmachine (3369, USA) at room temperature with a stretching rate of2 cm/min. At least 3 specimens of each membrane were tested andthe values were averaged.

.com

.cn

2.5. Proton conductivity and Methanol permeability

The proton conductivity of the circular samples (dimensions:1 cm) was measured by electrochemical impedance spectroscopyover the frequency range from 100 mHz to 100 kHz using anelectrochemical workstation (CHI660D). The scheme of the setupsof the through-plane conductivity is as described in Ref. [36]. Inaddition, the conductivity was measured at the relative humidity(RH) of 100% which was fixed for all the samples. The proton con-ductivity was calculated using the following formula: s¼ L·A�1 R�1,where L, A, and R are the membrane thickness, cross-sectional area,and resistance of the membrane, respectively.

The methanol diffusion coefficients were calculated throughformula variation of Fick's first law [37]: DK ¼ L $VB $CBðtÞ

A$CAðt e t0Þwhere CA

and CB are the methanol concentrations of methanol side (10 M)and water side, respectively; A, L, and VB are the effective area,thickness of the membrane, and volume of the permeatedcompartment, respectively; DK is the methanol permeability; t isthe end time, and t0 is the start time. Methanol concentration wasmonitored by gas chromatography (Agilent 7820) equipped with athermal conductivity detector (TCD) and a DB-624 column.

2.6. Single cell performance

5 mg/cm2 PteRu (1: 1) and 5 mg/cm2 PteC were coated on theanode and cathode of membrane by electrostatic painting,respectively. Carbon papers (model: Xinyuan power), which werepurchased from Kunshan Sunlaite Technology Limited (Jiangsu,China), were used as the gas diffusion layers. Then the single cellperformance of each MEA was characterized by a DMFC testingstationwith single serpentine flow fields (FCTS, Arbin Inc., America)at 60 �C under 100% RH. The flow rate of the feeding fuel (2 Mmethanol aqueous solution) was 2 mL min�1. The oxidant wasoxygen at a flow rate of 0.5 L min�1.

3. Results and discussions

3.1. Morphology of SPES-C and PVDF NFs

The SEM images of PVDF and SPES-C NFs are shown in Fig. 2(a)and (b). The corresponding diameter distribution maps are given inthe SI (Fig. S1). SPES-C NFs with diameters ranging from 70 nm to320 nm and PVDF NFs with diameters ranging from 50 nm to250 nm were arranged in a disorderly manner. Fig. 2 shows thatSPES-C and PVDF NFs were commonly curled in three dimensionsand closely entangled. This observation indicated that SPES-C andPVDF NFs were obviously different from the electrospun NFs.

3.2. Structure of SPES-C/PVDF composite membranes

Fig. 3(a)e3(c) show the surface images of different NCMs. All ofthe membranes were completely compact with SPES-C NFsstanding out of the PVDF matrix, which benefits the transport ofprotons. Fig. 3(a)e3(c) clearly show that the PVDF matrix encap-sulates SPES-C NFs in the membrane. Furthermore, we could alsodetermine the differences between each other. Fig. 3(a)e3(c) showthat the content of PVDF led to a considerable change on the surfaceof composite membranes. The NFs on the surface of themembranesbecame sparse as the content of the PVDF increased. This result wasconsistent with the increase in the matrix. In particular, numerousPVDF matrices encapsulated SPES-C NFs thoroughly. A dense cen-tral base polymer layer was formed in the internal membranes. Thislayer is shown in Fig. 3(d) and (e). The NFs were well dispersed inthe polymer matrix. The inner part showed a composite structurein which the NFs were dispersed in the PVDF matrix, thereby

w.spm

.com

.cn

Fig. 2. SEM images of solution-blown (a) PVDF and (b) SPES-C NFs.

Fig. 3. SEM images of surfaces: (a) SPES-C/PVDF-20, (b) SPES-C/PVDF-30, (c) SPES-C/PVDF-40. (d) and (e) SEM images of cross-section of SPES-C/PVDF-20 (red and green patternmarked as PVDF matrix and SPES-C NFs, respectively), (f) AFM phase image of surface of SPES-C/PVDF-40. (For interpretation of the references to colour in this figure legend, thereader is referred to the web version of this article.)

H. Wang et al. / Journal of Power Sources 365 (2017) 92e97 95

wwensuring the continuous channel for proton conduction. The PVDFformed a continuous methanol-resistant layer by flowing into thevoid spaces of SPES-C NFs. The sectional images of the differentlocations of NCMs showed different distribution directions of theNFs in the membrane. This observation revealed that the solution-blown NFs are commonly curled in three dimensions. This condi-tion is helpful to improve proton conductivity and dimensionalstability.

Fig. 3(f) shows that the AFM phase images in the tapping modeof membrane surfaces presented a bright/dark nanophase-separated morphology, with the dark regions representing thehydrophilic domains [38,39]. These domains absorbed water andfacilitated the transport of water and protons, whereas the brightregions corresponding to the hydrophobic domains providedmethanol resistance and dimensional stability in the membranes.The phenomenon also revealed that the hydrophilic phase domainswere apparently well connected to form proton channels, and theNFs were dispersed in the PVDF matrix to form dense membrane.

The XRD curves of all the samples are shown in Fig. 4(a). Thepeak of SPES-C was flat and occurred at 2q of 17.6�. The peak ofNCMs mainly occurred at 2q of 18.46�, 20�, 26.72�, and 39�, and thePVDF showed a wide dispersion peak around 21� which indicted alow degree of crystallinity. However, the peaks of NCMs remainedsharp at approximately 20�. The reason is that, after the hot-presstreatment, the NCMs showed a high degree of crystallinity, which

corresponded to 200/110 reflections of the b-phase, and weakpeaks at approximately 18.2� and 27.2�, which corresponded to 020and 111 reflections of the a-phase, respectively [25]. This phe-nomenon indicated that the PVDF in the internal membranes was acontinuous layer that could block methanol crossover. Besides, thepeaks for NCMs also showed flat with the increasing content ofSPES-C. This result was caused by the SPES-C redistribution in thePVDF. This result also revealed the good compatibility of thematerials.

3.3. Tensile strength

The mechanical properties of different membranes and SPES-CNFs are shown in Fig. 4(e). All composite membranes had excel-lent mechanical properties compared to PVDF membrane. More-over, the mechanical performance of the composite membranesinitially increased and then decreased with the increase in thePVDF content. This phenomenon occurred mainly because thematrix could not connect NFs when the content of PVDF was low.By contrast, when the content of PVDF was high, NFs could notreinforce the membranes well. Hot-press treatment could improvethe mechanical property of the PEMs [40]. Dimensional stability,oxidative stability and selectivity of different membranes areshown in SI.

www.spm

.com

.cn

Fig. 4. (a) XRD curves; (b) proton conductivity at varying temperatures under 100% RH; (c) methanol permeability; (d) single cell performances at 60 �C under 100% RH; and (e)mechanical properties of all the samples.

H. Wang et al. / Journal of Power Sources 365 (2017) 92e9796

3.4. Proton conductivity and Methanol permeability

Proton conductivity and methanol permeability of all the NCMsare summarized in Fig. 4(b) and (c), respectively. Proton conduc-tivity was measured at varying temperatures (i.e., 20 �Ce80 �C,100% RH). The plots of conductivity showed that the conductivity ofthe novel designed membranes exhibited positive dependencywith temperature. This finding revealed that proton conduction is athermally activated process. Fig. 4(b) shows that the proton

conductivity of SPES-C/PVDF-F was the lowest and close to 0.However, the proton conductivities of the membrane withthrough-membrane NFs were all higher than those of SPES-C/PVDF-F. These proton conductivities could obtain 0.1 S cm�1 at80 �C under 100% RH (SPES-C/PVDF-20). However, the protonconductivity of SPES hybrid membrane prepared by Wen was only0.047 S/cm under the same condition [41]. Therefore, the through-membrane and continuous SPES-C NFs extended from one side toanother side of the membrane can provide benefits. The proton

m

H. Wang et al. / Journal of Power Sources 365 (2017) 92e97 97

www.sp

conductivities decreased with the increase in the content of PVDF.This finding revealed that the decrease in the surface distribution ofthe NFs results in the gradual encapsulation of conduction channelsby the PVDFmatrix. Therefore, the existence of the “barrier layer” ofproton transportation at the surface of membranes hindered pro-ton transport and NFs can form proton-conducting channels in themembrane to improve the proton conductivity. The methanolpermeabilities of all the NCMs and Nafion were measured at roomtemperature. The methanol permeabilities of all the NCMs werevery low and could be as low as 1/500 that of Nafion117(14.1 � 10�7 cm2 s�1). The increase in the content of SPES-C NFsresulted in the increase in methanol permeability. The probablereason is that the methanol resistance layers of PVDF becameslightly efficient because of the decrease in the content of the PVDFmatrix. Therefore, the presence of PVDF successfully hindersmethanol crossover.

3.5. Single cell performance

Fig. 4(d) shows the polarization curves of the single cells fabri-cated by the composite membranes at 60 �C and 100% RH. The cellperformances of the composite membranes decreased with theincrease of PVDF contents. Furthermore, at the current density of100mA/cm2, the output voltages of SPES-C/PVDF-20, SPES-C/PVDF-30, and SPES-C/PVDF-40 are measured as 0.37692, 0.41538, and0.43385 V, respectively, while that of Nafion is only 0.35 V or lower[42e44]. All their sing cell performances have competitive advan-tage compared to Nafion.

4. Conclusion

In this study, we reported a novel approach to prepare com-posite PEMs with through-membrane and continuous nano-channels and methanol-blocking layer through SBP and hot-presstreatment. The existence of the “barrier layer” of proton trans-portation at the surface of membranes could hinder proton trans-port and NFs on surface can form proton-conducting channels inthe membrane improve the proton conductivity. The through-membrane NFs significantly improved the proton conductivityand cell performance of the PEMs. The mechanical performance ofcomposite membranes initially increased and then decreased withthe increase in the PVDF content. The composite membranes wereeffective in hindering methanol permeability because of the pres-ence of the continuous methanol-blocking layer, and the cell per-formance of composite membrane showed comparable resultscompared to Nafion. A substantial decrease in SR occurred. The SRsof all the composite membranes were less than 10%, and themembranes exhibited improved dimensional stability when SPES-CNF is incorporated into the PVDF matrix. The results revealed thatthe novel structure design can be considered a novel route toprepare high-performance and cost-effectivemembranes for DMFCapplications.

Acknowledgements

The author would like to thank National Natural Science Foun-dation of China (51473121), National Key R&D Plan(2016YFB0303304), the Science and Technology Plans of Tianjin(15PTSYJC00230, 13JCYBJC16700 and 14TXGCCX00014), Nationalscience and technology support program (2015BAE01B03), theFund Project for Transformation of Scientific and TechnologicalAchievements from Jiangsu Province (BA2015182), and the Programfor Changjiang Scholars and Innovative Research Team in Univer-sity (PCSIRT) of Ministry of Education of China (Grand no.

IRT13084) for their financial supports.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jpowsour.2017.08.072.

.com

.cn

References

[1] S.J. Peighambardoust, S. Rowshanzamir, M. Amjadi, Int. J. hydrogen Energy 35(2010) 9349e9384.

[2] V. Neburchilov, J. Martin, H. Wang, J. Zhang, J. Power Sources 169 (2007)221e238.

[3] S.G. Chalk, J.F. Miller, J. Power Sources 159 (2006) 73e80.[4] H. Li, G. Zhang, J. Wu, C. Zhao, Y. Zhang, K. Shao, M. Han, H. Lin, J. Zhu, H. Na,

J. Power Sources 195 (2010) 6443e6449.[5] K. Yoshimura, K. Iwasaki, Macromolecules 42 (2009) 9302e9306.[6] F. Zhang, N. Li, S. Zhang, S. Li, J. Power Sources 195 (2010) 2159e2165.[7] M. Divona, Z. Ahmed, S. Bellitto, A. Lenci, E. Traversa, S. Licoccia, J. Membr. Sci.

296 (2007) 156e161.[8] R.K. Nagarale, G.S. Gohil, V.K. Shahi, J. Membr. Sci. 280 (2006) 389e396.[9] I. Shabani, M.M. Hasani-Sadrabadi, V. Haddadi-Asl, M. Soleimani, J. Membr.

Sci. 368 (2011) 233e240.[10] H. Li, Y. Liu, J. Mater. Chem. A 2 (2014) 3783e3793.[11] S. Zhang, G. He, X. Gong, X. Zhu, X. Wu, X. Sun, X. Zhao, H. Li, J. Membr. Sci. 493

(2015) 58e65.[12] H. Wang, X. Li, X. Zhuang, B. Cheng, W. Wang, W. Kang, L. Shi, H. Li, J. Power

Sources 340 (2017) 201e209.[13] X. Xu, L. Li, H. Wang, X. Li, X. Zhuang, RSC Adv. 5 (2015) 4934e4940.[14] S. Shahgaldi, M. Ghasemi, W.R. Wan Daud, Z. Yaakob, M. Sedighi, J. Alam,

A.F. Ismail, Fuel Process Technol. 124 (2014) 290e295.[15] C. Lee, S.M. Jo, J. Choi, K. Baek, Y.B. Truong, I.L. Kyratzis, Y. Shul, J. Mater. Sci. 48

(2013) 3665e3671.[16] B. Zhang, X.P. Zhuang, B. Cheng, N. Wang, Y. Ni, Mater Lett. 115 (2014)

248e251.[17] R. Takemori, H. Kawakami, J. Power Sources 195 (2010) 5957e5961.[18] J.D. Snyder, Y.A. Elabd, J. Power Sources 186 (2009) 385e392.[19] R. Sood, S. Cavaliere, D.J. Jones, J. Rozi�ere, Nano Energy 26 (2016) 729e745.[20] J. Lee, N. Kim, M. Lee, S. Lee, J. Membr. Sci. 367 (2011) 265e272.[21] S. Yun, J. Woo, S. Seo, L. Wu, D. Wu, T. Xu, S. Moon, J. Membr. Sci. 367 (2011)

296e305.[22] U. Stachewicz, F. Modaresifar, R.J. Bailey, T. Peijs, A.H. Barber, ACS Appl. Mater.

Inter 4 (2012) 2577e2582.[23] D.M. Yu, S. Yoon, T. Kim, J.Y. Lee, J. Lee, Y.T. Hong, J. Membr. Sci. 446 (2013)

212e219.[24] M.M. Hasani-Sadrabadi, I. Shabani, M. Soleimani, H. Moaddel, J. Power Sources

196 (2011) 4599e4603.[25] L. Shi, X. Zhuang, X. Tao, B. Cheng, W. Kang, Fiber Polym. 14 (2013)

1485e1490.[26] S. Shi, X. Zhuang, B. Cheng, X. Wang, J. Mater. Chem. A 1 (2013) 13779.[27] X. Zhuang, K. Jia, B. Cheng, X. Feng, S. Shi, B. Zhang, Chem. Eng. J. 237 (2014)

308e311.[28] X. Zhuang, X. Yang, L. Shi, B. Cheng, K. Guan, W. Kang, Carbohyd. Polym. 90

(2012) 982e987.[29] H. Wang, X. Zhuang, J. Tong, X. Li, W. Wang, B. Cheng, Z. Cai, J. Appl. Polym. Sci.

132 (2015) 42843.[30] K. Chae, K. Kim, M. Choi, E. Yang, I.S. Kim, X. Ren, M. Lee, Chem. Eng. J. 254

(2014) 393e398.[31] S. Chul Gil, J. Chul Kim, D. Ahn, J. Jang, H. Kim, J. Chul Jung, S. Lim, D. Jung,

W. Lee, J. Membr. Sci. 417-418 (2012) 2e9.[32] B. Decker, C. Hartmann-Thompson, P.I. Carver, S.E. Keinath, P.R. Santurri,

Chem. Mater 22 (2010) 942e948.[33] L. Jia, X. Xu, H. Zhang, J. Xu, J. Polym. Sci. B Polym. Phys. 35 (1997) 2133e2140.[34] M.S. Subramanian, G. Sasikumar, J. Appl. Polym. Sci. 117 (2010) 801e808.[35] S. Ren, G. Sun, C. Li, Z. Wu, W. Jin, W. Chen, Q. Xin, X. Yang, Mater. Lett. 60

(2006) 44e47.[36] T.D.O. Gadim, F.J.A. Loureiro, C. Vilela, N. Rosero-Navarro, A.J.D. Silvestre,

C.S.R. Freire, F.M.L. Figueiredo, Electrochim. Acta 233 (2017) 52e61.[37] J. Zhang, Z. Zhou, Chin. J. Power Sources 31 (2007) 721e724.[38] W.H. Lee, K.H. Lee, D.W. Shin, D.S. Hwang, N.R. Kang, D.H. Cho, J.H. Kim,

Y.M. Lee, J. Power Sources 282 (2015) 211e222.[39] D.W. Shin, S.Y. Lee, N.R. Kang, K.H. Lee, D.H. Cho, M.J. Lee, Y.M. Lee, K.D. Suh,

Int. J. Hydrogen Energ 39 (2014) 4459e4467.[40] J.B. Ballengee, P.N. Pintauro, J. Membr. Sci. 442 (2013) 187e195.[41] S. Wen, C. Gong, Y. Shu, F. Tsai, J. Yeh, J. Appl. Polym. Sci. 123 (2012) 646e656.[42] H. Lin, S. Wang, J. Membr. Sci. 452 (2014) 253e262.[43] H.S. Thiam, W.R.W. Daud, S.K. Kamarudin, A.B. Mohamad, A.A.H. Kadhum,

K.S. Loh, E.H. Majlan, Energ. Convers. Manag. 75 (2013) 718e726.[44] H. Ahmad, S.K. Kamarudin, U.A. Hasran, W.R.W. Daud, Int. J. Hydrogen Energ

36 (2011) 14668e14677.